Cooperation in the Prisoner's Dilemma with Anonymous Random Matching

Abstract

The paper considers the repeated prisoner's dilemma in a large-population random-matching setting where players are unable to recognize their opponents. Despite the informational restrictions cooperation is still a sequential equilibrium supported by “contagious” punishments. The equilibrium does not require excessive patience, and contrary to previous thought, need not be extraordinarily fragile. It is robust to the introduction of small amounts of noise and remains nearly efficient. Extensions are discussed to models with heterogeneous rates of time preference and without public randomizations.